Since there is great uncertainty regarding the true prevalence of people with FASD in the general population, and the prevalence of people with FASD who abuse or are addicted to substances, specifically, in combination with the lack of FASD-specific data pertaining to specialized addiction treatment services in Canada, several important assumptions were used and sensitivity analyses were conducted, as described step-by-step below.
Source of data
Data were obtained from the Drug and Alcohol Treatment Information System (DATIS; http://www.datis.ca), which monitors the use of specialized addiction treatment services by people in Ontario, Canada. Started in 1992, DATIS collects data on the numbers and types of clients entering publicly funded specialized addiction treatment across the province. This system includes approximately 200 treatment programs, administered by 170 agencies [23, 24]. The types of services provided by each agency can vary; some agencies provide a specific type of service (e.g., assessment and referral, withdrawal management, individual counseling), while others provide a comprehensive mix of services. Regardless of the type of service, all services are delivered free-of-charge to Ontario residents (i.e., they are covered by the province’s universal health insurance plan). Agency-level participation has been over 95% since 2000.
The DATIS database is structured by admissions to treatment programs, such that a new admission is triggered when a client enters a new treatment program or transitions between two different types of services. Since 2002, data entry has been supported by a web-based user platform accessible by all frontline clinicians working in the designated programs across the province. Data entry fields correspond to 66 data elements that are mandatory for all admissions. Sociodemographic characteristics, information on substance use, and other treatment-related factors (e.g., referral source, current and past diagnoses of mental disorders, and treatment mandates) are entered by the service provider at admission, typically following the first face-to-face encounter with the client. Unique identifiers for individual clients and agencies are generated automatically by the software, as is a variable documenting the type of treatment program or service (i.e., outpatient, residential or residential withdrawal management). Details on service use, including the number of outpatient visits and days or residential care, are entered by service providers at discharge. Data are stored on a central server located at the Centre for Addiction and Mental Health in Toronto, Canada.
All admissions corresponding to services received during the 2010/11 fiscal year (April 1, 2010 to March 31, 2011; N=91,333) were extracted for analysis. This included all admissions occurring during the fiscal year, as well as those where treatment had started prior to April 1, 2010, but that had continued into the study period. As noted above, an admission corresponds to a particular type of service, with movement between types of services (e.g., from a residential service to an outpatient program) counted as separate admissions. Multiple admissions per individual client were included. This was done to ensure that the most complete estimates of the volume and costs of services delivered during the study period were obtained. Therefore, it is important to note that each admission does not necessarily represent a separate individual, as one individual can have multiple admissions.
In order to determine whether or not a client has had a “lifetime mental disorder” (i.e., a diagnosed mental disorder at any point within their lifetime; response: yes versus no), the question “Have you ever been diagnosed by a qualified mental health professional with a mental disorder within the last 12 months or within your lifetime?” was asked to each client at the beginning of treatment.
Age groups, divided into 5-year intervals from 14 years of age and younger to 70 years of age and older, were generated from the clients’ date of birth. All variables (lifetime mental disorder, date of birth, and sex) were self-reported by the clients during their initial clinical encounter.
Service type (i.e., outpatient, residential treatment, or residential withdrawal management, automatically generated in DATIS) was abstracted, as were the number of visits for outpatient treatment and days in residential treatment and residential withdrawal management. The numbers of visits and days of care, entered by service providers based on client charts, provide estimates of the volume of services received. For admissions that began prior to or ended after the 2010/11 fiscal year, only those visits/days that occurred within the study period were counted. To comply with Personal Health Information Protection Act (2004), cells with values less than 6 were redacted and replaced with "<6".
Estimation of the prevalence of clients with FASD in specialized addiction treatment and their utilization of these services
Number of people with FASD in Canada
The prevalence of FAS and FASD are currently unknown in Canada. However, the most commonly cited rough estimates are 1 per 1,000 for FAS , and 9 per 1,000 for FASD . Using data on the general population, by age group and sex, of Canada in 2010 , and assuming a prevalence of 9 per 1,000 for FASD , the number of people with FASD in Canada was estimated.
Number of people with FASD who abuse or are addicted to alcohol and/or drugs
In order to estimate the number of clients with FASD who had received specialized addiction treatment services in Canada in 2010/11, the prevalence of individuals with FASD who abuse or are addicted to alcohol and/or drugs was calculated based on available epidemiological studies [7–10] using the meta-analysis technique described below.
Prevalence data from the epidemiological studies concerning alcohol/drug abuse/use/dependence among those with FASD were transformed into log-odds for the meta-analysis . Log-odds estimates were weighted by the inverse variance of the log-odds transformed prevalence. Heterogeneity between studies was assessed using the Cochrane Q-test and the I2 statistic [29, 30]. The prevalence estimates were pooled using the Mantel-Haenszel method, using a random-effects model .
Publication bias was tested by: 1) visually inspecting a funnel plot for skewed distribution, 2) using a ranked correlation test , and 3) employing a weighted regression test . Publication bias was then adjusted for using the trim and fill method .
Number of people with FASD who utilized specialized addiction treatment services and their rate of utilization in 2010/11
In order to calculate the number of people with FASD who utilized specialized addiction treatment services in 2010/11, it was assumed that the rate of specialized addiction treatment services among this population (individuals with FASD) was the same as the rate among individuals with a lifetime mental disorder. This assumption is based on a very high prevalence of co-morbid mental illness reported among individuals with FASD [7, 9, 35, 36]. Based on the authors’ comprehensive literature review (Popova et al., unpublished), the weighted mean for mental retardation (International Classification of Diseases, version 10 [ICD-10] category: F70-F79) among individuals with FASD is 48% (95% confidential interval [CI]: 44.4%-51.4%) and for disorders of psychological development (ICD-10 category: F80-F89) is 37% (95% CI: 35.9%-39.0%).
In order to estimate the rate of utilization of specialized addiction treatment services for individuals with a lifetime mental disorder, the number of specialized addiction treatment services admissions for individuals with a lifetime mental disorder was divided by the total number of individuals with a mental illness in Ontario (reported by the Ministry of Health and Long-Term Care  and Health Canada ).
In turn, in order to estimate the total number of admissions, and visits/days, by treatment type, among clients with FASD for all of Canada, the distribution for each treatment type among clients who received specialized addiction treatment services in the province of Ontario was used and extrapolated to the total Canadian population. This approach is justifiable given that Ontario represents about 39% of the total population of Canada.
Estimation of costs
The cost for specialized outpatient treatment ranged from $60 to $109 per service and for residential treatment ranged from $138 to $314 per resident day in Canada in 2010/11 (Martin et al., in progress). These unit costs are estimated based on the costs reported by five Local Health Integration Networks (LHINs) across Ontario. The overall estimates are inclusive of the cost of supervision, facility costs, salaries, and other sundry expenses. The ranges reported account for the differences in costs incurred due to the number of spots/beds available (capacity), whether the treatment is hospital based or community based, the intensity of activities provided, and staff professionalism (which affects both their salaries and the scope of the staff complement).
The corresponding costs - for outpatient treatment: $60 and $109 per service as the lower and upper estimates, respectively; and for residential treatment: $138 and $314 per resident day as the lower and upper estimates, respectively - were applied to substance-attributable specialized outpatient visits and residential days, in order to obtain the total costs of such services for clients with FASD.
All cost figures are presented in Canadian dollars.
Due to a very limited number of existing epidemiological studies, there is great uncertainty regarding the prevalence of individuals with FASD who abuse or are addicted to alcohol and/or drugs. As described above, the weighted mean of 37% (CI: 21.6%-54.5%), which was calculated based on the available epidemiological studies [7–10] was used in the main analysis. In addition, two separate analyses were performed assuming that 22% (as the lower estimate) and 55% (as the upper estimate; both are based on the estimated CI) of individuals with FASD abuse or are addicted to alcohol and/or drugs.